ChangeLog

2012-05-30  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/main.cu (getUsage): Removed optional parameters that were unnecessary.

2012-05-09  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* README: Updated. 

2011-11-17  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* test/testSearch.cu: Added test cases for searchQueries3() function.
	* src/main.cu: Moved NUM_THREADS macro from within search.cu to common.h.

2011-08-11  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/align.cu (getNumQrsPerIter_paired): Fixed a bug. 

2011-08-10  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/hitList.h: Removed 'seqOffset' data member because it wasn't being 
	used. 
	* src/refPosList.h: Removed 'seqOffset' data member because it wasn't 
	being used.

2011-07-22  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/smithWaterman.cu (computeMaxScore): Changed nucleotide comparison
	code such that the nucleotide 'N' matches any nucleotide.  

2011-07-05  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/reference.cu (insertionSort3_gpu): Added new function. 

2011-06-20  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/reference.cu (quickSort2_gpu): Added new function. 

2011-06-16  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/align.cu: Some query names in the latter half of the alignment 
	output file were strange. This is now fixed.
	* src/refPosMap.cu (binarySearch): Added new function. 

2011-06-15  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/search.cu (searchQueries_gpu): Speeded-up the concatenation of 
	query sequences. 

2011-05-24  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* test/testReference.cu: Added test cases to test strncpy_gpu function.
	* src/reference.cu (strncpy_gpu): Added function to copy characters 
	from one string to another.

2011-05-23  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* test/testReference.cu: Added test cases for getHash_gpu, 
	bubbleSort, arrSearch_gpu, getRandNum, arrGetRandomNums_gpu, getBestMatches, 
	and refSearchQuery_kernel.
	* src/reference.cu: Added functions to run on GPU such as getHash_gpu, 
	bubbleSort, arrSearch_gpu, getRandNum, arrGetRandomNums_gpu, getBestMatches, 
	and refSearchQuery_kernel.

2011-05-16  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/main.cu (getParameters): Added a command-line parameter that 
	indicates whether the program should be run on CPU only. 
	* src/output.cu (printOutput_paired): Added function to output aligned 
	paired-end queries. 
	* src/align.cu: Added paired-end alignment support.
	(getNumQueriesPerIteration_paired): Added function to count the number 
	of queries per iteration for paired-end alignment.
	* test/testSearch.cu: Added test cases for searchQueries_paired function.

2011-05-11  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* test/testReference.cu: Modified test cases.
	* src/reference.cu (refSearchQuery_paired): Modified code so that 
	result is passed as a parameter instead of being returned.
	(getHits): Modified code. 

2011-05-10  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* test/testReference.cu: Added test cases for refSearchQuery_paired 
	function.
	* src/reference.cu (refSearchQuery_paired): Added code for paired-end 
	alignment support.
	(getHits): Added code for paired-end alignment support.

2011-05-05  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* Makefile: Source is now first compiled into object files and 
	then the executable is created.
	* src/main.cu: Updated getParameters function to add support for 
	fetching the paired query sequence file from the command-line.
	Updated getUsage function to include parameter for fetching 
	paired query sequence file.
	* test/testQuery.cu: Added a test case for qryGetNext2 function.
	* src/query.cu: Added a function to get the next Query given the query 
	sequence file and a Query object.

2011-05-04  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* Makefile: Changed program name from 'gpuSW' to 'swift'.
	* README: Changed the program name from 'gpuSW' to 'swift'.
	* test/testSmithWaterman.cu: Added a test case to check for correct 
	behavior when the reference length is more than the number of bases 
	in the reference sequence.
	* src/smithWaterman.cu: Fixed a bug in computeMaxScore_cpu in which 
	the alignment was being performed even beyond the null character 
	of the reference sequence.

2011-05-03  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/smithWaterman.cu: Fixed a bug in which gap open and extension 
	penalties were being used in an incorrect manner. 
	

2011-04-26  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* test/testSmithWaterman.cu: Added test case for CPU-based 
	Smith-Waterman implementation.
	* src/smithWaterman.cu: Added CPU-based implementation of Smith-Waterman.

2011-04-21  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/main.cu: Gap opening and extension penalty was incorrectly 
	assigned to 'mismatch'. It's now corrected.
	Added command-line parameter "-v" to allow users to find out program 
	version.

2011-04-20  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/smithWaterman.cu: If a gap is openend on the query, isRefGapOpened
	is set to 0; if a gap is opened on the reference, isQryGapOpened is 
	set to 0. 

2011-04-19  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/smithWaterman.cu: Removed functions that were not being used.
	* test/testReference.cu (searchQuery): Now passing the max hits 
	parameter to the refSearchQuery function. 
	* test/testSearch.cu (searchQuery): Now passing the max hits parameter to 
	the searchQueries function.  
	* src/reference.cu (refSearchQuery): Now accepts as a parameter the max
	number of hits per query to be returned.
	(getBestScoringPosition): Now accepts as a parameter the max
	number of hits per query to be returned.
	* src/search.cu (searchQueries): Now accepts as a parameter the max
	number of hits per query to be returned.
	* src/main.cu (main): Now passing the max hits parameter to the 
	searchQueries function. 

2011-04-18  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* test/testSmithWaterman.cu (backtrack): Fixed a small bug.

2011-04-15  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* test/testSmithWaterman.cu: Added new test file.
	* test/testAlign.cu: Added new test file.

2011-04-14  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* test/testHitList.cu (duplicateNode): Added test case.
	* src/hitList.cu (hitListDuplicateNode): Added function.
	* src/smithWaterman.cu: Modified code so that now I don't get 
	"Unspecified launch failure" anymore. 

2011-03-31  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* test/testSearch.cu (searchQuery): Added test cases.
	* src/reference.cu (getBestScoringPosition): When the given list had 
	only one node, a pointer to the list was being returned instead of 
	a pointer to the copy of the list. This has been fixed.

2011-03-30  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/reference.cu: Removed refGetHashTable function.
	Removed getUniqueHashCount function.
	* src/smithWaterman.cu: Removed smithWaterman5 function.
	Removed smithWaterman4 function.
	Removed smithWaterman3 function.
	Removed smithWaterman2 function.
	Removed computeMaxScore2 function.
	Removed computeMaxScore function.
	Removed backtrack function.
	Removed getPreviousCell function.
	Removed getArrayMax function. 
	Removed getBlockId function.
	Removed getThreadId function.
	Removed getSimilarityScore function.
	(backtrack): Renamed backtrack2 function to backtrack.
	(computeMaxScore): Renamed computeMaxScore2 function to computeMaxScore.
	* src/memory.cu: Removed allocateDeviceMem2 function.
	Removed allocateHostMem2 function.
	Removed copyMemDeviceToHost2 function.
	* src/input.cu: Removed adjustSeqOffset function.
	Removed getSequences function.
	Removed getSequences2 function.
	Removed getSequences3 function.
	* src/hitList.cu: Removed bubbleSort function.
	* src/align.cu: Removed adjustSeqOffset2 function.
	Removed getOffsets2 function.
	Removed getQuerySeq function.
	Removed getRefSeq function.

2011-03-29  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/reference.cu (refSearchQuery): Fixed bug where reference sequence 
	fragment offset was set to a strange value when subtracted by 
	a number bigger than the fragment offset.
	Removed getAvlblHashTablesFile function.
	Removed refGetHashTableFile function. 
	Removed refPrintHashTable function.
	Removed refUpdateAvlblHashTablesFile function.
	Removed avlblHashTablesFile variable.

2011-03-28  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/input.cu (getSeqName): Changed fileOffset type from int to ulong.
	(getSeq): Changed fileOffset type from int to ulong.
	* src/reference.cu (refSearchQuery): Reference sequence fragment offset 
	is set to reference sequence offset if the fragment offset is less 
	than the sequence offset.

2011-03-24  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/align.cu (getMemPerRefQryPair): Removed H matrix from memory 
	requirement calculation.
	* src/smithWaterman.cu (smithWaterman): Improved performance and 
	fixed bugs.

2011-03-21  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/reference.cu (refPreprocess): Not caching the reference hash anymore 
	because I don't see any performance benefits in maintaining a cache.
	(getBestScoringPosition): Changed array allocation of clusterSizeArr 
	from stack to heap because stack based array allocation was throwing 
	segmentation fault for some queries. 

2011-03-18  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* Makefile: Removed list.cu.
	* src/query.cu: Removed qryPrintMatches and qryPrintMatches2 functions 
	from query.cu.
	Removed qryInitializeMatchFile because it was no longer needed.
	* src/preprocess.cu: Removed preprocess and preprocessRef functions 
	because they were not needed anymore.
	* src/search.cu (printHeuristicMatches): Made function static.
	* test/testReference.cu (preprocess): Added unit test for refPreprocess
	function.
	* test/testRefPosList.cu (pushBack): Added test.
	(isEqual): Added unit test for refPosListIsEqual function.
	* src/refPosList.cu (refPosListPushBack): Added function.
	(refPosListIsEqual): Added function.
	Removed refPosListAppendUnique function because it was not needed anymore.
	* README: Added instruction on profiling the code.

2011-03-17  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/hitList.cu (merge): Now using hitListPushBack instead of 
	hitListJoin function.
	* src/reference.cu (refGetHashTable): Hash table is now initialized to 
	NULL.
	(refPreprocess): Modified function.
	(refSearchQuery): Modified function.
	(refSearchQuery): Now using hitListPushBack instead of hitListJoin function.
	(getBestScoringPosition): Now using hitListPushBack instead of hitListJoin 
	function.
	* test/testReference.cu (searchQuery): Added test cases.

2011-03-15  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/reference.cu (refSearchQuery): Made the query parameter constant.
	* test/testQuery.cu (create): Added test function.
	* src/query.cu (qryCreate): Added new function.

2011-03-14  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/refPosList.cu (refPosListDelete): Fixed a bug.
	* test/testHitList.cu (popFront): Added test.
	* src/hitList.cu (hitListPopFront): Added function.
	(mergeSort): Added merge sort function.
	* src/reference.cu (getBestScoringPosition): Added code to free up 
	heap memory.
	(refSearchQuery): Added code to free up heap memory.
	* src/search.cu (searchQueries): Once a filtered results file is printed in 
	phase-1, list of HitList nodes are deleted.

2011-03-11  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/reference.cu (getBestScoringPosition): Made this function static.
	(refSearchQuery): Changed the order in which lists to be joined are passed 
	to hitListJoin function.
	* test/testHitList.cu (join): Added test cases for hitListJoin function.
	(createNode): Added check for 'next' and 'prev' properties of the 
	created node.
	(getSize): Added test cases for hitListGetSize function.
	(copyNode): Added test case for hitListCopyNode function.
	* src/hitList.cu (hitListCmpNodes): Added fumnction to compare nodes.
	(bubbleSort): Transferred node comparison code to hitListCmpNodes function.
	(hitListJoin): Modified code to set 'prev' property of appended list.
	(hitListCopyNode): hitListCopyNode function now returns a pointer to the 
	destination node.

2011-03-10  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* test/testHitList.cu (sort): Added test for hitListSort function.
	(swapAdjNodes): Removed test for hitListSwapNodes and added test for 
	hitListSwapAdjNodes function.
	(cmpNodes): Added test cases for hitListCmpNodes function.
	* src/hitList.cu (hitListSwapAdjNodes): Modified hitListSwapNodes function 
	to hitListSwapAdjNodes function.
	* src/reference.cu: Removed refSearchQuery2 function.

2011-03-09  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* README: Updated description, hardware and software requirements, and 
	installation.
	* src/main.cu (getParameters): If one of the required command-line 
	parameters are not entered during program invocation, the program 
	prints usage and exits.
	* src/input.cu (getOffsets): Added an additional parameter to determine 
	whether the offsets correspond to the reverse complement of a query.
	* test/testInput.cu (getOffsets): Added a new test.

2011-03-08  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/align.cu (alignQueries): Calling adjustSeqOffset function to adjust 
	the reference sequence offset.
	* src/reference.cu (refPreprocess): File offset of a sequence fragment 
	is now the same as the file offset of the whole sequence.
	* src/input.cu (adjustSeqOffset): Added function. 
	* test/testInput.cu: Added new test file. 
	Added a few test cases.

2011-03-07  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/preprocess.cu: Added explicit casting in the pow functions.
	* src/reference.cu (refPreprocess): Added explicit casting in the pow
	function.
	* src/align.cu (alignQueries): Changed how reference sequence offset is 
	calculated.
	* src/input.cu (getSeq): Breaking out of the loop when a new sequence 
	ID is encountered.
	(getOffsets): Instead of using shift, now using offset as the reference 
	distance.
	(getOffsets): Changing back to using shift instead of offset as the 
	reference distance.
	(getOffsets): Added a condition such that if reference distance is less 
	than 0, it is set to 0.
	* src/output.cu (printOutputSam): The output file pointer is now being 
	passed as a parameter instead of the file name.
	(printOutput): The output file pointer is now being passed as a 
	parameter instead of the file name.

2011-03-04  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/array.cu: Added file.
	* test/testArray.cu (arraySuite): Added file.
	* src/reference.cu (getRandomNums): Added function.

2011-03-03  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/output.cu (printOutput): Output will now be printed to a file.
	(printOutputSam): Output will now be printed to a file.
	(printCigar): Output will now be printed to a file.
	* src/align.cu (alignQueries): Added match score, mismatch score and 
	output file as parameters.
	* src/main.cu (getParameters): Added new parameters.
	Made the function static.
	(getUsage): Modified.
	(main): Passed match score, mismatch score and output file as parameters 
	to alignQueries function.
	* src/common.h: Added new macros.
	* test/testReference.cu (getHashTableFile): Added new test cases.
	* src/reference.cu: Created new functions and modified existing ones.
	(refGetBestScoringPosition): Returning multiple best-matches instead of 
	only 1.
	(refPreprocess): Only non-overlapping tuples of the reference are 
	considered.

2011-03-01  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* test/testReference.cu (isHashTableAvailable): Added 2 test cases.
	* src/reference.cu (refPrintHashTable): Added function.
	* README: Added file.
	* src/main.cu (main): Now passing the match file as an argument to 
	alignQueries and searchQueries functions.
	(main): Now passing the output format as an argument to alignQueries. 
	(getParameters): Now returns a 1 on success and 0 on failure.
	(getUsage): Added program usage.
	(main): Number of references and queries and the size of reference 
	and query are now automatically calculated. 
	* src/search.cu (searchQueries): Match file is now passed as a parameter.
	* src/align.cu (alignQueries): Match file is now passed as a parameter.
	(alignQueries): Output format is now passed as a parameter.

2011-02-28  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* Makefile: Changed the name of the executable.
	* src/reference.cu (refPreprocess): Fixed reference name and sequence 
	offset calculation.
	(refPreprocess): Fixed reference sequence offset calculation. 
	* src/query.cu (qryGetNext): Fixed query sequence offset calculation.

2011-02-25  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/input.cu (getOffsets): Modified code so that input can be parsed 
	in the new format. 
	* src/search.cu (printHeuristicMatches): Modified code so that query name 
	offset, query sequence offset, reference name offset, and reference 
	sequence offset can be printed to the intermediate output file.
	* src/query.cu (qryGetNext): Added code to calculate nameOffset and 
	seqOffset.
	(qryGetNext): Reverse complement is now created in this function.
	(qryGetNext): Removed unnecessary variables.
	* src/query.h: Added new data members.
	Added new data member.
	* test/testMain.cu (main): Added test suite for RefPosList.
	* test/testRefPosList.cu: New source file.
	(create): Created new test case.
	(create): Added new test.
	(deleteList): Added new test.
	* test/testRefPosList.h: New header file.
	* src/refPosList.h: Removed refPosListSort function.
	Removed bubbleSort function.
	Added new data members to structure.
	* src/refPosList.cu: Removed refPosListSort function.
	Removed bubbleSort function.
	(refPosListPrint): Now also prints new data members.
	(refPosListCreate): Created new function.
	(refPosListDelete): Created new function.
	* src/reference.cu: Removed refGetNext function.
	Removed refGetFileId function.
	Removed refAddFile function.
	Removed fileArr and fileArrSize variables.
	Removed refGetNameId function.
	Removed refAddName function.
	Removed refArr and refArrSize variables.
	Removed refGetPseudoPosition function.
	Removed refGetNumPartitions function.
	Removed refGetPartitionIndex function.
	(refInitialize): Removed memory allocation for ref->seq.
	Removed memory allocation for ref and ref->name.
	(refPreprocess): Added code to calculate nameOffset and seqOffset of 
	the reference sequence.
	(refPreprocess): Changed how new node is appended to lookupTable.
	(refSearchQuery): Changed the interface.
	* src/reference.h: Removed refGetNext function.
	Removed refGetFileId function.
	Removed refAddFile function.
	Removed refGetNameId function.
	Removed refAddName function.
	Removed refGetPseudoPosition function.
	Removed refGetNumPartitions function.
	Removed refGetPartitionIndex function.
	* src/hitList.cu (hitListSwapNodes): New function.
	(bubbleSort): Now using hitListSwapNodes function to swap nodes.
	(hitListPrintNodes): Corrected printf format specifier for 'unsinged 
	long int'.
	(hitListSwapNodes): Removed warning message.
	* src/hitList.h (hitListSwapNodes): New function.
	Added 'static' modifier to bubbleSort function. 
	* test/testHitList.cu: Added test cases to test hitListSwapNodes function. 

2011-02-24  Pankaj Gupta  <pankaj.gupta@stjude.org>

	* src/hitList.h: Changed hitListCreateNode function declaration 
	to accomodate new data members.
	* src/hitList.cu: Added additional parameters to hitListCreateNode function.
	(hitListPrintNodes): Changed to accomodate new data members.
	(hitListCopyNode): Changed to accomodate new data members.
	* src/common.h: Added more macros.
	* src/search.cu (printHeuristicMatches): New function.